arXiv subjects

Dakui Wang

Publications and source records attributed to Dakui Wang.

5 recordsLinked to original sources

arXiv2026

Self-Authored Verification Is Unreliable in Heuristic Self-Improving Agents

Self-improving agents accumulate capability by repeatedly rewriting procedural policies, controllers, or heuristic rules. They typically rely on self-authored tests or metrics to decide whether to accept subsequent edits. The agent controls both the optimized object and its verifier. As a result, self-assigned scores can remain near perfect while real deployment performance degrades or stays low. We study this problem through the verifier--deployment gap. This gap refers to the discrepancy between an agent's self-authored verification signal and a sealed deployment evaluation that the agent cannot observe or access. We ask how self-authored verification fails under iterative policy-and-test rewriting, how the failure changes with capability, and how little exogenous trust is sufficient to prevent real regressions from being deployed. To address this problem, we introduce a Sealed Exogenous Acceptance Loop (SEAL). SEAL retains self-authored tests but compares each candidate with the incumbent through a fixed harness-side audit. The agent cannot author or inspect the audit, receives only accept/reject, and the whole incumbent state is retained after a clear regression. Our experiments show that this problem often appears in heuristic learning settings. These settings require trial-and-error discovery of the target objective. We further find that failures of self-written verification are stratified by capability. Weaker agents tend to damage previously acquired strategies behind easy self-tests. Stronger agents are more stable, but they still mismeasure the deployment distribution. Standard self-written constraints do not reliably close this gap. In contrast, SEAL outperforms unprotected baselines across six models and three random seeds. Reliable self-improvement need not abandon self-verification, but it requires at least one deployment-acceptance signal outside the agent's control.

cs.CL↗

arXiv2025

From Large to Super-Tiny: End-to-End Optimization for Cost-Efficient LLMs

Large Language Models (LLMs) have significantly advanced artificial intelligence by optimizing traditional Natural Language Processing (NLP) workflows, facilitating their integration into various systems. Many such NLP systems, including ours, directly incorporate LLMs. However, this approach either results in expensive costs or yields suboptimal performance after fine-tuning. In this paper, we introduce a three-stage cost-efficient end-to-end LLM deployment pipeline, comprising prototyping, knowledge transfer, and model compression, to effectively tackle the cost-performance dilemma in LLM-based frameworks. Its high cost-efficiency is manifested not only in simplifying system complexity and producing super-tiny online models with enhanced performance and reduced costs in the results, but also in addressing development cycle constraints, the lack of extensive high-quality data, and limited computational resources during the project development process. In the first stage, we construct an optimal performance prototype system by transforming complex tasks into a function call-based LLM-driven pipeline, which serves as a teacher model to generate high-quality data. In the second stage, we combine techniques like rejection sampling fine-tuning, reinforcement learning, and knowledge distillation to transfer knowledge to 0.5B student models, delivering effective performance at minimal cost. In the final stage, we further compress models to 0.4B via quantization and pruning, achieving ultra-low latency and cost. Extensive experimental results and the framework's modular design suggest cross-domain capabilities and potential applicability in other NLP areas.

cs.CL↗

arXiv2024

Can Multimodal Large Language Model Think Analogically?

Analogical reasoning, particularly in multimodal contexts, is the foundation of human perception and creativity. Multimodal Large Language Model (MLLM) has recently sparked considerable discussion due to its emergent capabilities. In this paper, we delve into the multimodal analogical reasoning capability of MLLM. Specifically, we explore two facets: \textit{MLLM as an explainer} and \textit{MLLM as a predictor}. In \textit{MLLM as an explainer}, we primarily focus on whether MLLM can deeply comprehend multimodal analogical reasoning problems. We propose a unified prompt template and a method for harnessing the comprehension capabilities of MLLM to augment existing models. In \textit{MLLM as a predictor}, we aim to determine whether MLLM can directly solve multimodal analogical reasoning problems. The experiments show that our approach outperforms existing methods on popular datasets, providing preliminary evidence for the analogical reasoning capability of MLLM.

cs.CL↗

arXiv2021

Efficient Byzantine-Resilient Stochastic Gradient Desce

Distributed Learning often suffers from Byzantine failures, and there have been a number of works studying the problem of distributed stochastic optimization under Byzantine failures, where only a portion of workers, instead of all the workers in a distributed learning system, compute stochastic gradients at each iteration. These methods, albeit workable under Byzantine failures, have the shortcomings of either a sub-optimal convergence rate or high computation cost. To this end, we propose a new Byzantine-resilient stochastic gradient descent algorithm (BrSGD for short) which is provably robust against Byzantine failures. BrSGD obtains the optimal statistical performance and efficient computation simultaneously. In particular, BrSGD can achieve an order-optimal statistical error rate for strongly convex loss functions. The computation complexity of BrSGD is O(md), where d is the model dimension and m is the number of machines. Experimental results show that BrSGD can obtain competitive results compared with non-Byzantine machines in terms of effectiveness and convergence.

cs.DC↗

arXiv2015

Comment on "Clustering by fast search and find of density peaks"

In [1], a clustering algorithm was given to find the centers of clusters quickly. However, the accuracy of this algorithm heavily depend on the threshold value of d-c. Furthermore, [1] has not provided any efficient way to select the threshold value of d-c, that is, one can have to estimate the value of d_c depend on one's subjective experience. In this paper, based on the data field [2], we propose a new way to automatically extract the threshold value of d_c from the original data set by using the potential entropy of data field. For any data set to be clustered, the most reasonable value of d_c can be objectively calculated from the data set by using our proposed method. The same experiments in [1] are redone with our proposed method on the same experimental data set used in [1], the results of which shows that the problem to calculate the threshold value of d_c in [1] has been solved by using our method.

cs.LG↗